论文部分内容阅读
基于变尺度混沌优化的方法可以利用混沌变量的特定内在随机性和遍历性来跳出局部最优点,并可以变尺度搜索提高局部空间的搜索速度和精度。把该方法应用到神经网络的权值优化中,可以得到很好的效果。
The method based on variable scale chaos optimization can make use of the specific internal randomness and ergodicity of chaotic variables to jump out of local optimal points and to improve the search speed and accuracy of local space by changing the scale. The method is applied to the neural network weight optimization, you can get good results.